1. Technical Field
The present disclosure concerns a computer-readable recording medium for recording an object from a captured image and providing information associated with the recognized object.
2. Description of the Related Art
In image object recognition, a finger object is relatively easily recognized because of its substantially regular shape in comparison to other recognition targets. However, to identify a finger, initially feature point comparison is typically performed for various sizes and angles. As an example of finger recognition, after a finger shape is stored as a template, a finger shape may be recognized using the stored template. Further, to improve recognition rate, skin color may be used. In response to finger recognition, a recognized finger path is traced to recognize a moving path of a fingertip. A known Lucas-Kanade (LK) scheme for recognizing a position based on a size of a finger and a change in an angle may be used.
However, in finger-recognition-based input, operations are performed in involving different variables (for example, various sizes and angles) of a stored finger template. Use of these variables improves recognition rate but increases computation required. Since computation needs to be performed for the regions of a captured image, complex computation is required for a high-definition image.